Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#42546 - [systemd] Unable to Open systemd --user Session: pam_systemd connection times out repeatedly
Attached to Project:
Arch Linux
Opened by Alexander Stein (ajstein) - Saturday, 25 October 2014, 10:07 GMT
Last edited by Dave Reisner (falconindy) - Sunday, 22 March 2015, 15:11 GMT
Opened by Alexander Stein (ajstein) - Saturday, 25 October 2014, 10:07 GMT
Last edited by Dave Reisner (falconindy) - Sunday, 22 March 2015, 15:11 GMT
|
DetailsDescription:
It appears after running a series of updates on 22 October 2014 (see attached pacman log), I am no longer able to get a functional systemd --user session. Foolishly or not, I have continued to check pacman and make sure I am up to date. Nonetheless, after login on tty1 (my default) there is a long delay. Checking the journal, I always see the following. Oct 22 18:28:34 mbp62 login[894]: pam_systemd(login:session): pam-systemd initializing Oct 22 18:28:34 mbp62 login[894]: pam_systemd(login:session): Asking logind to create session: uid=1000 pid=894 service=login type=tty class=user desktop= seat= vtnr=0 tty=tty1 display= remote=no remote_user= remote_host= Oct 22 18:28:34 mbp62 systemd[1059]: pam_systemd(systemd-user:session): pam-systemd initializing Oct 22 18:28:59 mbp62 login[894]: pam_systemd(login:session): Failed to create session: Connection timed out I have tried adding debug flags to /etc/pam.d/ after reading the man pages, but I cannot get further information. I am not sure how to troubleshoot further. I have read only a few BBS postings and one bug report where you recommended cleaning the system. I am not sure if it is the same issue or not yet. Additional info: * package version(s) http://dl.il5.in/pacman.log * config and/or log files etc. The last month or two of /usr/bin/login messages showing this pattern crop up starting 22 October. http://dl.il5.in/journal.login.failures.log My full journal from systemd for this boot. http://dl.il5.in/journal.this.boot.log Steps to reproduce: 1 - Boot computer 2 - Unlock dmcrypt/luks disks 3 - Wait for boot to complete 4 - Login with non-root user on tty1 (my default) 5 - Switch to tty12 and watch journal log on this terminal (per my settings) 6 - Notice error 7 - Observe my terminal sessions report IPC/Dbus errors because last bashrc command systemctl --user import-environment fail. NOTE: I inquired about this on BBS (https://bbs.archlinux.org/viewtopic.php?pid=1468535#p1468535) and IRC and no one seemed able to help me with this. I apologize in advance if this is a foolish configuration error. |
This task depends upon
Closed by Dave Reisner (falconindy)
Sunday, 22 March 2015, 15:11 GMT
Reason for closing: No response
Additional comments about closing: Not likely to be a packaging problem. The user manager still has some warts until kdbus is available.
Sunday, 22 March 2015, 15:11 GMT
Reason for closing: No response
Additional comments about closing: Not likely to be a packaging problem. The user manager still has some warts until kdbus is available.
http://dl.il5.in/etc.systemd.tar.xz
http://dl.il5.in/systemd.user.configs.xz
I have tried with and without loginctl enable-linger username and disable-linger. I was using enable-linger as this bug happened, and then I turned it off as part of testing and this did not impact the remaining timeout problem.
[falconindy]
Server = http://repo.falconindy.com/$repo/os/$arch
And then attach to it with gdb and get a backtrace.
(gdb) bt
#0 0x00007f843083a9d3 in __epoll_wait_nocancel () from /usr/lib/libc.so.6
#1 0x00007f843138833d in sd_event_run (e=0x7f8432c7f240, timeout=<optimized out>) at src/libsystemd/sd-event/sd-event.c:2262
#2 0x00007f8431352c32 in manager_run (m=0x7f8432c7e010) at src/login/logind.c:1162
#3 main (argc=<optimized out>, argv=<optimized out>) at src/login/logind.c:1223
How do I get the full backtrace, and log it to a file to send to you better yet?
Oct 25 19:36:24 mbp62 systemd-logind[2235]: New seat seat0.
Oct 25 19:36:24 mbp62 systemd-logind[2235]: New session c3 of user al.
Oct 25 19:36:24 mbp62 systemd-logind[2235]: New session c1 of user al.
Oct 25 19:39:19 mbp62 systemd-logind[2235]: Removed session c3.
Oct 25 19:39:36 mbp62 systemd-logind[2235]: New session c2 of user al.
Oct 25 19:41:09 mbp62 systemd-logind[2235]: kernel does not support evdev-revocation
Should I be worried about this?
If you want more debugging information from systemd itself, I'd suggest:
# systemd-analyze set-log-level debug
So I guess I should reboot, log in as root, attach to logind with gdb, and then login with the user with the bad sysd --user sessions? How do I get a running full backtrace logged to a file?
Will post back shortly after I see systemd-analyze set-log-level debug. Expect a response back in the next 15-20 minutes.
http://dl.il5.in/dmesg.log
http://dl.il5.in/gdb-systemd-logind-533.log